Over the last 18 months, the material advancement pixel-based tracking and attribution capabilities have lifted podcast advertising to new and exciting heights. This has allowed Ad Results Media (ARM) to continuously evolve and offer profoundly more insightful analysis and reporting regarding the interaction effect of when a listener downloads a podcast episode to that same listener’s visit our client’s website.
Although we still use and triangulate data related promo code redemptions, vanity URL web visits, and/or post-checkout survey results, the proliferation of pixel data has opened up our ability to measure and analyze a much more granular set of listener engagement data (web impressions, unique visitors, revenue, etc…) for media planning and campaign optimizations.
Recently, we analyzed listener/customer IP data for a client to determine the zip codes of those pixeled listeners who downloaded a podcast episode that contained the client’s ad and subsequently visited our client’s website. We then married that data with data from the most recent U.S. Census results to generate an estimate of the average household income for each show/listener/client combination (results shown below). As you’ll see, the Avg HH income across the entire data set is $87,360. This is quite a bit higher than the 2020 national average of $67,521 and tells us that the listeners who engaged with the ads were likely more affluent and thus had more disposable income than the mean. It validated our media plan and targeting strategy as this product is a luxury apparel item.
Networks will frequently provide survey data as to what the HH income of their shows is. This data is an aggregate of the show’s audience, which can create some false signals. By looking at the zip code level of individuals who downloaded the podcast episodes, we can obtain a much more accurate assessment of the show’s audience, whether it matches our client’s target audience and how it impacts performance.
To further illustrate this point, take the podcast, “Ratchet & Respectable”. The high average HH income is likely due to the fact that they’re based in Brooklyn, NY, and their audience over-indexes in high-income earners from Wall Street. Thus, if you are an advertiser looking to reach this demographic, geo, or combination thereof, the data provided from this level of granular analysis is valuable in creating efficiencies in your media plan.
Another interesting analysis we ran for the same client was comparing the percentage of listeners/downloaders by state (which we get from the zip codes), with the percentage that population represents in the US.
(Note: Attribution percent is the percent of listeners/downloaders from the specified region, and US percent is the percent population those listeners represent out of the US. The State Index is the formula of “Attribution Percent/US Percent - 1” which we can use to find markets that over-index in listenership)
What immediately popped out was the number of regions that significantly over-indexed in listeners/downloaders compared to the population in the U.S. D.C. emerged as the top state (district), however, we also saw that VT, OR, WA, MA, CO, ID, ME, MN, MT were over 100% represented. This validated the performance we were seeing for the client as they sell luxury apparel made for cold weather. So, intuitively, the direction of the over-performance made sense to us.
Additionally, the data’s insights created opportunities to spend efficiently in other advertising channels. For example, the data suggested that this client should test a local-radio campaign in the promising markets listed in the data. The data also suggested they should test a programmatic/DI strategy for specific geographies.
Pixel-based attribution data on its face provides very interesting information compared to traditional tracking methods. However, diving a bit deeper with secondary information (such as the zip codes of the downloaded IP addresses) provides the opportunity to grow and scale our client’s campaigns in incredibly efficient and effective ways.